[This paper studies a stylized model in which a division manager's historical cost reports are verified. In a one-period setting, the problem of tacit collusion between the verifier and the division manager is so severe that no mechanism can be constructed under which there is a unique equilibrium that has the verifier exercising anything other than the minimal level of care. However, by extending the contractual relationship to two periods, the tacit collusion problem can be resolved. A feature of the optimal collusion-preventing long-term contract is that it relies on history-contingent production decisions.]

[This paper studies information system design in a model of double moral hazard in which there is both a decision problem and a control problem. If either problem is considered in isolation, an information system that provides more public information is preferred. However, an information system that provides less public information can, in fact, be desirable because of an interaction between the two problems. The benefit of choosing an information system that provides less information is that it serves as a substitute for commitment for the principal. The cost is that neither the principal's decision (act) nor the agent's payments can be conditioned on the information. We provide sufficient conditions under which less information and more information are each optimal.]

This paper develops a positive role for accounting conservatism in fostering relational contracts between two agents in a two-period model of moral hazard. Building on Kreps (1996), the principal in our model designs a conservative measurement system and optimal contracts to create multiple equilibria that foster a team-based corporate culture. Accruals introduced by conservatism increase each agent's stake in the future of the relationship when it matters most—when it is going badly. This makes staying in the relationship worthwhile for the agents, even if they plan to play a low payoff equilibrium in the second period to punish first-period free-riding. In turn, this allows the principal to use lower-powered (and less costly) team incentives in the first period of the relationship. In contrast, deferred compensation increases each agent's stake in the future of the relationship when it is going well, making it less efficient in fostering relationships.

We study a principal‐agent model in which there is an option to defer a capital project approval decision. A control (incentive) problem makes the option to wait valuable when it would not have been valuable otherwise. Deferring the project approval decision has both a cost and a benefit. The cost of waiting is that the agent’s uncertainty regarding future project cost realizations cannot be exploited. However, by delaying the first project’s approval decision, the principal can condition its approval on the agent’s cost report of the second project. Such conditioning can be valuable in the provision of incentives because of a diversification effect.

In a team setting, wherein only group performance is tracked, we show that muted incentive contracts may be sufficient to motivate team members. By having the team repeat a task, explicit (contractual) incentives can be substituted by implicit incentives team members provide to each other. We also study an example in which, despite uncorrelated individual performance measures being available, it is optimal to condition each manager's pay on both managers' performance. This can be viewed as creating a group performance measure. Using a group performance measure provides each manager with incentives to monitor and a means of punishing other managers.

ABSTRACT This article studies contracts between a principal and an agent that are robust to information asymmetries about measurement quality. Our main result is that an information asymmetry about measurement quality not only reduces the usefulness of a given performance measure for stewardship purposes, it also qualitatively changes the way the performance measure is used if the information asymmetry is sufficiently large. We also study the manipulability of performance measures, assuming that poor measurement quality creates room for manipulation via selective (cherry‐picked) corrections by the agent. With known imperfect measurement quality, manipulability lowers the cost of providing incentives. Manipulability introduces overstatements only, while imperfect measurement introduces both overstatements and understatements. However, with an information asymmetry about measurement quality, manipulability can increase the cost of providing incentives, since there is now an induced information asymmetry about manipulability.